Fundamentals of Machine Learning Courseware (55375AC)

This five-day course offers an introduction to the fundamental principles of machine learning classification models. It covers the creation of these models, the importance of data quality, and the process of testing and validating models. The course is platform-agnostic and requires a basic familiarity with Python and a foundational understanding of differential calculus, probability, and statistics.

This course provides a comprehensive introduction to machine learning models, with a particular focus on classification models. The course begins with an overview of machine learning and its various models, including classification, clustering, and regression. Students will learn about the purpose of these models and the types of problems they can solve.

The course then delves into the specifics of different algorithms used to create classification machine learning models. It covers a range of algorithms, including Decision Trees, Random Forests, Gradient Boosted Trees, XGBoost, Logistic Regression, and the K-Nearest Neighbors Algorithm, among others. The course provides a detailed comparison of these algorithms, discussing their strengths, weaknesses, and appropriate use cases.

The process of creating a classification model is covered in depth, with a focus on data preparation, model construction and tuning, and testing and validation. The course also explores the differences between binary and multi-valued classification and demonstrates how to create a multi-class classification model.

The course also includes a review of key statistical concepts and techniques used to analyze the distribution, scale, and relationships between items in a dataset. This knowledge is crucial for understanding the validity of a machine learning model.

The course then guides students through the process of refining a machine learning model by selecting the most relevant features from the dataset, examining the distribution of values, investigating correlation between features, normalizing data, and removing bias.

The final modules of the course focus on measuring the performance of a classification model, understanding regularization to reduce overfitting, and evaluating a model. The course also discusses the challenges of using an imbalanced dataset to create a classification model, how to recognize potential problems, and how to address them.

The course spans over five days and is intended for developers and analysts who are new to machine learning and want to understand how machine learning classification models work. It requires some familiarity with Python, an understanding of matrix and vector arithmetic, and some basic familiarity with probability, statistics, and differential calculus.